CLIMATE CHARACTERIZATION AND HISTORICAL CHANGES IN DENSITY AND INTENSITY OF LIGHTNING AROUND THE 500 KV BACABEIRA-PARNAÍBA TRANSMISSION LINE

CLIMATE CHARACTERIZATION AND HISTORICAL CHANGES IN DENSITY AND INTENSITY OF LIGHTNING AROUND THE 500 KV BACABEIRA-PARNAÍBA TRANSMISSION LINE Fernando DINIZ*¹, Ana Clara MARQUES; Pedro REGOTO; Luciano RITTER; Thiago FERREIRA, Rafael ALIPIO, Euro ALMEIDA, William MEJIA, Fabian ROJAS, Oscar GONZALEZ ¹Argo Energia Brazil [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], wm[email protected], [email protected], [email protected] SUMMARY This study investigates the impact of lightning on the operation of transmission lines (TLs), particularly the 500 kV Bacabeira-Parnaíba TL in Northeast Brazil, in the context of climate change. The research aims to understand the changes in lightning behavior over time, as variations in thunderstorm frequency and intensity can affect TL efficiency and safety. The Bacabeira-Parnaíba TL, spanning 300 km with 566 towers, was segmented into sections of approximately 25 x 20 km for analysis. Using data from Earth Networks from 2013 to 2023, the study conducted monthly and annual analyses of lightning density and intensity. Findings indicate an increasing trend in lightning density and intensity along the TL, suggesting more frequent and severe thunderstorms. Comparison with data from the Brazilian National Electric System Operator (ONS) shows similar patterns in lightning density around the TL for the period 1998-2013. However, recent data (2016-2023) reveal higher lightning densities, particularly in the TL's western part. The study also compares two approaches for calculating the TL's lightning performance: the traditional method using an average lightning density and a detailed method using data from Earth Networks. This comparison aims to provide insights into the accuracy of current lightning performance calculations for TLs. The findings highlight the need to consider changing lightning patterns in the planning and operation of transmission lines to ensure safety and efficiency. KEYWORDS Climate change, Lightning, Transmission Line 10953 C4 POWER SYSTEM TECHNICAL PERFORMANCE PS3 - Insulation Co-Ordination and Lightning Interference Analysis: Challenges, Opportunities and Advances 2024 Paris Session 1 1 Introduction The basic network transmission lines (TLs), which transport energy from the North and Northeast regions to the Southeast of Brazil, suffer from changes in climatic patterns due to physical changes in the Earth's system. In general, one of the main impacts on TLs is related to the change in the frequency and intensity of severe storms associated with lightning discharges. In recent years, the Brazilian energy transmission sector has experienced more occurrences of interruptions and power outages. In the current context of complementarity of energy sources, new installations of wind and solar farms need to somehow drain/transport their energies, which in this case are via transmission lines. In other words, TLs will be increasingly required for the continuity of Brazilian energy expansion. Therefore, the objective of this work is to understand the seasonal behavior of the occurrence of cloud-to-ground lightning discharges and whether or not there has been a change in the pattern of density over the surroundings of the 500-kV Bacabeira-Parnaíba transmission line, located between the states of Maranhão and Piauí, in the Northeast region of Brazil. The relevant questions that will be answered in this work are the following: A) What is the preferred region for the highest occurrence of electrical discharges over the 500kV Bacabeira-Parnaíba transmission line? B) What is the period/year with the highest electrical activity over the analyzed region? C) What is the month/period of the year in which the highest electrical activity occurs over the analyzed region? D) Is there any pattern of changes over the years in electrical activity in the study area? E) What is the impact of the aforementioned factors on the calculation of transmission line performance? 2 Study Area The study area is the surroundings of the 500-kV Bacabeira-Parnaíba transmission line under concession by Argo concessionaire, located between the states of Piauí and Maranhão, in the Northeast region of Brazil (Figure 1). The TL was divided into similar parts, from Tower 1 to Tower 566. The divisions were composed of an area of approximately 25 x 20 km rectangle. From the division made, the changes in the characteristics of the lightning strikes along the entire length of the transmission line can be more visually and practically compared. Figure 1: Area around the 500-kV Bacabeira-Parnaíba transmission line under the concession of Argo. The polygons represent the divisions made, of the areas individually analyzed of approximately 25 x 20 km. 2 3 Data and Methods 3.1 Datasets The data used for lightning discharges were from Earth Networks, which is a global network that has been improving its detection of cloud-to-ground and cloud-to-cloud lightning yearly. The sensors have a broadband system with detection in the frequency range between 1 and 12 MHz. This allows the sensors to detect not only strong cloud-to-ground lightning but also weak cloud-to-cloud pulses. The sensor records the entire waveform of each lightning strike and sends it in compressed packet format to a central server. Instead of only using peak pulses, the full wave signal is used to locate lightning and distinguish between cloud-to-cloud and cloud-to-ground. This more detailed information improves the efficiency of detection and the accuracy of the system's location. New digital signal processing technologies are used on the server to ensure high-quality detections and eliminate false locations. When lightning occurs, electromagnetic energy is emitted in all directions. All Earth Networks sensors that detect the waveforms register and send those waves to the central server via the internet. The precise arrival times are calculated by correlating the waveforms of all sensors that detected the pulses of a lightning strike. The arrival time of the wave and the amplitude of the signal can be used to determine the electric current peak of the lightning strike and its exact location, including latitude, longitude, and altitude. It is worth mentioning the uncertainties associated with lightning records by the Earth Networks sensor network, so that the data used in this study are as accurate as possible and are used in the best way. The study by Zhu et al. (2017) evaluated the performance of the Earth Networks network with data observed from the Lightning Observatory in Gainesville and Camp Blanding, Florida, United States, for the years 2014 and 2015. The authors found an average error of about 200 meters for the location and around 15% for the electrical intensity (amplitude of the lightning current). They emphasized that lightning strikes with higher current peaks had a higher probability of detection and correct classification. The increase in the number of sensors over the past few years has resulted in improved detection of lightning discharges, both in the number of lightning strikes and in the accuracy of location. Therefore, it is important to assess the annual change in the record of electrical discharges considering the error associated with this measurement and the increase in the number of sensors year over year. For this study, the quantities used were the total annual density and the monthly average density. The total density consists of the monthly and annual frequency of lightning strikes. The data has a 1 km spatial resolution and the period used for the analyses is from January 1, 2013 to June 30, 2023. Although the year 2023 is analyzed only with the data from the first half, the wettest period and with the highest number of lightning strikes typically comprises the months between the end of summer and autumn in the region analyzed. Therefore, it is possible to infer that the first half of 2023 encompasses the majority of the lightning strikes expected for the year 2023. 3.1 Methods The present study has two types of analysis. The first one is the spatial analysis of the density of lightning discharges. For these, annual lightning density maps were constructed, where all the lightning strikes observed at each grid point are accumulated annually and the areas with the highest detection density and lowest density are checked. This same analysis was also carried out for monthly data, where the lightning strikes recorded in all months, for example for January 2013 to 2023, were accumulated for each grid point and then divided by the number of years, in this case, 11 years. Thus, the average monthly density of lightning discharges is obtained. The second type of analysis developed in this study is in heatmap format. For this, the lightning density of each grid point within each of the 11 divisions of 25 x 20 km around the TL was calculated. We performed a spatial mean within each rectangle division. This gives the average density in each rectangle evaluated, for each year from 2013 to 2023. Finally, annual lightning density maps are created for the periods 2013 to 2023 and 2016 to 2023. These maps are used to compare with data from the National Electric System Operator (ONS in 3 Portuguese), the institute responsible for coordinating and controlling the operation of generation and transmission facilities of electrical energy in Brazil. 4 Results - Annual Overview In this section, annual lightning data are evaluated spatially and tabularly. Analyses of cloudto-ground lightning density allow for the interpretation of historical changes throughout the vicinity of the 500-kV Bacabeira-Parnaíba transmission line, for the period from January 1, 2013, to June 30, 2023. 4.1 Spatial Analysis Map analyses help to detect the densities of lightning strikes to understand the signs of changes in the variables determined throughout the study area. For the analyses of the maps shown below, the blue colors indicate a lower frequency of electrical discharges (lightning density), while the warm colors (yellow-orange-red) indicate a higher frequency. Figure 2 shows the behavior of the total annual cloud-to-ground lightning density for the period 2013-2023 around the 500-kV Bacabeira-Parnaíba transmission line. In the years 2013 to 2017, the lightning density was significantly lower than in the later years of the analysis (2018-2023). The year 2023, despite being considered a shorter period (from January to July), was the one with the highest occurrences around the transmission line. The western portion of the line (T1 to T101) was the preferred region for the highest concentration of lightning strikes in the analyses. However, in 2020, the eastern portion of the line (T501-T566) showed a higher concentration of lightning strikes. This shows that there is a greater occurrence of cloud-to-ground lightning at the ends of the line than in the central part. Figure 2: Total annual cloud-to-ground lightning density around the 500-kV Bacabeira-Parnaíba transmission line, for the period 2013-2023. 4.2 Heatmap Analysis The following analysis presents the total annual density of cloud-to-ground lightning in each division of the 500 kV Bacabeira-Parnaíba TL. In this way, the analyses make it possible to interpret the changes in the 2013-2023 period in a regionalized point of view, as indicated in the study area (Figure 1). For the following figures, the yellow-orange-red colors indicate lower density, while the purple colors indicate higher values. 4 Figure 3 shows the behavior of the spatial average of the total annual cloud-to-ground lightning density for the period 2013-2023 for each division carried out in this work for the 500-kV BacabeiraParnaíba TL. In 2020, 2021, 2022, and 2023, divisions T1 to T51 and T51-T101 had a significantly higher lightning density than the other regions, exceeding 6.4 occurrences per km². From 2018 onwards, there was a clear increase in the total lightning density in the areas around the 500-kV Bacabeira-Parnaíba TL in all the towers, especially between 2020 and 2023. In 2021, there were 8.5 lightning strikes per km² in the second division (T1-T51). On the other hand, in 2020, in the eastern divisions, between T501 and T566, there was a more significant increase in lightning strikes (approximately 5.2 strikes per km²), which was not observed in the other years (maximum of 4.4 strikes per km²). The year 2023 stands out with the highest density of lightning strikes recorded on most of the transmission line. The first two divisions analyzed (T1 to T101) had the highest lightning density of any series since 2013. Figure 3: Spatial average of the total annual lightning density for each division of the 500-kV Bacabeira-Parnaíba TL, for the period 2013-2023. 5 Results - Monthly Overview This section assesses the monthly behavior of the average cloud-to-ground lightning density around the 500 kV Bacabeira-Parnaíba transmission line. The analyses are made to evaluate the main period of lightning occurrence throughout the year. Figure 4 shows the behavior of the average monthly density (January to December) of cloudto-ground lightning for the period 2013-2023 around the 500 kV Bacabeira-Parnaíba TL. For the analysis of the maps shown below, the blue colors indicate a lower frequency of electrical discharges (lightning density), while the warm colors (yellow-orange-red) indicate a higher frequency. The period with the highest lightning density throughout the year is from January to April. November and December are transition months, when the amount of lightning gradually increases, 5 especially to the west of the TL. May, June, and July mark the transition to a drier period with less lightning throughout the year. In August, September, and October, the average density of lightning strikes is low, indicating that lightning-generating events are infrequent at this time of year. Figure 4: Average monthly cloud-to-ground lightning density around the 500 kV Bacabeira-Parnaíba TL, for the period 2013-2023. The annual average from January to June covers the period from 2013 to 2023, while the months from July to December are calculated only for the period from 2013 to 2022. 6 ONS Comparison Figures 5 and 7 show the average annual cloud-to-ground lightning density around the 500 kV Bacabeira-Parnaíba transmission line for the 2013-2023 and 2016-2023 periods, respectively. The analysis is made in comparison with the average annual cloud-to-ground lightning density data provided by the National Electricity System Operator (ONS in Portuguese) [3], shown in Figure 6. The comparison between the annual lightning density for 2013-2023 drawn up in this study (data from Earth Networks) and the annual lightning density for 1998-2013 from ONS (data from the Tropical Measuring Mission - TRMM satellite) indicates a similar pattern. Thus, the two data sources present annual lightning density values that are relatively close. In addition to the analysis of the annual average for the 2013-2023 period (Figure 5), we present a map of the annual average for 2016-2023 (Figure 7). When we compare the two periods, it is clear that the average for the last 8 years has a higher density of lightning strikes compared to the entire period (2013-2023), which may be associated with an increase in the occurrence of lightning strikes over the last few years. When comparing the annual average for the most recent period (2016-2023) with data from the ONS (1998-2013), we note that the density of lightning strikes recorded over the last eight years has been higher, especially to the west of the 500 kV Bacabeira-Parnaíba transmission line. The values exceed 7 lightning strikes/km²/year between T1 and T51, T51 and T101 and T101 and T151. In the area around towers T51 to T101, there are maximum values above 9 lightning/km²/year. 6 Figure 5: Average annual lightning density around the 500 kV Bacabeira-Parnaíba transmission line, for the period 2013-2023. Figure 6: Average annual lightning density around the 500 kV Bacabeira-Parnaíba transmission line for the period 1998-2013. The data comes from the TRMM satellite, with a spatial resolution of 25 km x 25 km. Modified from ONS (https://www.ons.org.br/paginas/sobre-o-sin/mapas). 7 Figure 7: Average annual lightning density around the 500 kV Bacabeira-Parnaíba TL, for the period 2016-2023. 7 Impact of Historical Ground Flash Density Changes on 500-kV Bacabeira-Parnaíba TLs Lightning Performance The results shown in the previous sections have a direct impact on the calculation of the outage rate of the transmission line. This section explores some of these impacts but does not intend to exhaust the subject. 7.1 Methodology for Calculating the Performance of the TL The 500-kV Bacabeira-Parnaíba TLs predominantly feature guyed suspension structures along their route, with an average nominal effective height of 42.1 m and an average span of 530 m. The geometry of the typical guyed tower is presented in Figure 8, along with the main data of the TL conductors. The sag of the phase and shield wires are, respectively, 22.86 m and 19.40 m, based on the mechanical load definitions on the cables. Phase conductors: CAL 1120 1010 kcmil (diameter 29.34 mm – RCC=0.06797 Ω (60ºC)) 4 conductors per phase spaced 900 mm apart Shield wires (SW): 1) EHS 3/8'' (diameter 9.52 mm – RCC=4.0538 Ω (40ºC)) 2) OPWG (diameter 13.6 mm – RCC=0.60264 Ω (40ºC)) Cartesian coordinates of the conductors on the tower (in meters): Phase A (-7.3, 42.55); Phase B (0, 50.05); Phase C (7.3, 42.55); SW1 (-6.2, 54.51); SW2 (6.2, 54.51) Figure 8: Geometry of the typical guyed tower and data of the TL conductors. 8 To assess the lightning performance of the 500-kV Bacabeira-Parnaíba TLs, the transmission system components were modelled in the Alternative Transients Program (ATP). The models adopted follow the latest recommendations of CIGRE [4]. The tower was modelled as a single-phase distributed line, with its surge impedance calculated according to its geometrical data. Similarly, the guy wires were modelled with surge impedances based on their geometric characteristics. The simulations assume direct lightning strikes to the top of a central tower in the system, considering five spans on each side of the strike point. Each span is represented as an untransposed line section with distributed/frequency-dependent parameters. The grounding of the tower footing is concisely modelled with a variable resistance value to cover different grounding conditions of the structure. The injected current corresponds to a double-peak waveform, accurately replicating the main median parameters of currents measured by Berger [5] at Mount San Salvatore. In the performed simulations, the peak value of the injected current waveform is varied, and the integration method is applied to the overvoltages impinging across the line insulators to determine whether insulation breakdown will occur [6]. The peak value of the lightning current that causes insulation breakdown, leading to line outage, is called the critical current 𝐼𝐶. The line backflashover rate (BFR) for a given section 𝑘 of the line, is then computed as: 𝐵𝐹𝑅𝑘= 0.6 × 𝑁𝑔𝑘 × 𝐾𝑔𝑒𝑜 × 𝑃(𝐼𝑃> 𝐼𝐶) (1) where 𝑃(𝐼𝑃> 𝐼𝐶) is the probability of the lightning peak current exceeding the minimum current that causes insulation breakdown, the factor 0.6 is used to discount the effect of strokes along the span, 𝑁𝑔𝑘 is the ground flash density of the line section 𝑘, and 𝐾𝑔𝑒𝑜 is a factor associated with the exposure area of the TL, which primarily depends on the average height of the TL structures. 7.2 Results Considering the models described in the previous section, the representative transmission system of the 500-kV Bacabeira-Parnaíba TLs was implemented in ATP. For each specific grounding condition (tower-foot resistance varied between 10 Ω and 60 Ω), the overvoltages per kA on the line's insulator strings in response to lightning strikes were simulated. Then, the integration method was applied to determine the peak value of the lightning current that would cause insulation breakdown, and thereafter, the probability of this critical current being exceeded [4]. Finally, using equation (1), it is then possible to generate curves that relate the TL's outage rate to the tower-foot grounding resistance. Figure 9 illustrates the obtained results. Each curve represents the outage rate calculation for a given ground flash density (Ng) as follows: The black curve was computed assuming an average Ng along the entire TL, quantified as 2.2 from Figure 3. The other four curves were derived considering the TLs divided into four sections, namely T1-T101, T101-T151, T151-T401, and T401-T566, with average Ng values equal to 4.0, 2.4, 1.5, and 2.0, respectively, computed from Figure 3. The horizontal line in the figure indicates the maximum outage rate established by ONS for 500 kV TLs in Brazil.